Ultrasensitive Fiber-Optic Sensor for AI-Enhanced Voice Recognition and 360° Acoustic Detection

Fiber-optic sensors offer distinct advantages for acoustic signal detection under extreme conditions due to their immunity to electromagnetic interference and capability for remote sensing. However, limitations in sensitivity, response time, and directionality have restricted their broader application in real-world scenarios. In this work, we demonstrate an ultrasensitive fiber-optic sensor featuring a record-low detection limit of 0.4 mPa and a rapid response time of 6 μs for acoustic detection. This sensor is based on a microfiber coupler that provides an exceptionally broad acoustic response range, spanning from 20 Hz to 25 kHz, with a frequency resolution of 0.04 Hz at 200Hz. Leveraging these capabilities, the sensor achieves a remarkable 97.1% accuracy in intelligent voice recognition when integrated with deep learning algorithms. Additionally, the sensor is equipped to recognize directional acoustic signals and enables 360° full-range acoustic source localization with a dual-sensor configuration. This versatile sensor holds substantial potential for practical applications in human-machine interfaces and sonar positioning systems.